Lightspeed Attenuator a new passive preamp

[jkeny]

Gareth, Maybe something like this

being done here: http://diyparadise.com/web/index.php?option=com_content&task=view&id=143&Itemid=26

 

[Tolu]

I have been critical, too. But that Lightspeed device is the best preamp or better volume control I have ever heard!

 

[Bigun]

I'm getting quite enthusiastic about this idea.

Jkeny - the issue remains matching for 6 channels (5 surrounds and 1 sub). The paradise web site finishes with this

"Please be patient. We are still working on our LDR solution. It has remote control and LED display as well. We just need to fix the channel mismatch problem."

I'm thinking to stick with Shunt only, no series LDRs. This halves the number.

Perhaps with my HT application the matching is not as critical compared with Stereo. Firstly, the sub is certainly not critical. So now I only have 5 channels to match. The rear-to-front balance is also non-critical so that means I can match the front 3 channels and back 2 channels somewhat independently. I suspect the match between centre and the front L/R isn't critical either. So perhaps I can boil this down to something I can manage ?

 

[Tolu]

-

 

[Tolu]

Hi Gareth

shunt only sounds worse in relation to series/shunt! I have tried it with the same goal than you. I have used a non-magnetic CMY55 Dale Vishay resistor but it isn't as good as LDR-series/shunt.

Do you think that you hear the benefits of the LDR solution with a HT setup?

Can you post the link of the paradise website?

 

[mikelm]

In my experience you do need a good quality resistor if you use series resistor / shunt if you want to match quality of series LDR, shunt LDR but it is possible.

I'm using Caddock TF020 as series R and I don't think I loose any quality

mike

 

[tubo]

Hi Tubo, was wondering how you got your Lightspeed Attenuator? TIA

dogears|Moises

hi dogears,

i got mine from allied electronics here in the US! sorry i forgot to update my profile...

 

[jkeny]

In my experience you do need a good quality resistor if you use series resistor / shunt if you want to match quality of series LDR, shunt LDR but it is possible.

I'm using Caddock TF020 as series R and I don't think I loose any quality

mike

That's interesting Mikelm & Tolu - a good series resistor is as good sounding as the series/shunt LDR - so this implies that the
resistor part of the LDR is as good as some top quality resistors. Puts pay to the idea that these LDRs have highish distortion? Or does it mean that we don't know the distortion spectra of the resistors that we use in audio. Very few resistor manufacturers give these distortion graphs.

Bigun, have you just joined the SKA forum? Matching is always going to be a problem unless/until some automated (microprocessor based?) solution is created

 

[Bigun]

you think that you hear the benefits of the LDR solution with a HT setup?

Can you post the link of the paradise website?

I've no idea if LDR is needed for HT. But I do need a volume control. A multi-ganged pot isn't cheap. DigiKey sell the Vishay P9A series that can be ganged up to 7 levels and uses plastic conductors (the P9S cermet seems impossible to find). But you have to buy several of them to build a multi-ganged unit. If you already have a psu in your box for the LDR approach it may not be that much more expensive - it all depends on the yield for matching.

My HT system is most likely over-engineered with regards quality of sound. If I were more pragmatic I would either buy a 5 channel amp off-the-shelf or build something less critical. But after enjoying the learning process of building the speakers I thought it would be interesting to use the HT project as an 'excuse' to learn about amplifiers. So far I am finding it very interesting!


p.s. the Paradise link is at the bottom of Jkeny's post (just a couple back from here)

 

[mikelm]

I think you will hear the difference with your HT system

 

[mikelm]

That's interesting Mikelm & Tolu - a good series resistor is as good sounding as the series/shunt LDR - so this implies that the
resistor part of the LDR is as good as some top quality resistors. Puts pay to the idea that these LDRs have highish distortion? Or does it mean that we don't know the distortion spectra of the resistors that we use in audio. Very few resistor manufacturers give these distortion graphs.

Yes I think the resistance quality is excellent . . . . but I am talking subjectively. . . perhaps some H2 etc but I think noise is very low and frequency response is very good.

 

[TJF]

Remember: Pietjers had tested the quality of serial resistors and is still using Shinkoh Tantalum (in his serial r./shunt LDR version), i.e. here

Something, what isn't still handled in this thread: What about constant input impedance at all volume steps and potential dividers? With one (and the same) serial resistance at all volume steps?

Remember: George used serial r. and shunt LDR in his MK1. MK2 points in the right direction ...

Regards
Thomas

 

[Bigun]

If the source doesn't mind driving the LDR input and the amplifier has a high enough input impedance, I can't see how a varying input impedance will be detrimental. But of course, somebody will need to do a listening test to confirm.

For the HT application, series-shunt is simply impractical because of matching.

 

[TJF]

But of course, somebody will need to do a listening test to confirm

That is always the key... first of all, if you do regulations, which are based on chance. Like (complete) different input impedances at different volume steps...

 

[mikelm]

But of course, somebody will need to do a listening test to confirm.

I did these tests - and for my set up I had no problem.

But my setup may not be typical

my series resistor is 2.2K so my CD o/p stages can in theory see very different resistances - from just over 2.2K to about 100k

for me it's fine because my CD o/p stage can handle this easily without any audible worsening of the sound.

I cannot imagine a 10K series resistor would create a problem for output stage so I really cannot see problem here unless you think the LDR type resistor sounds better that the resistor you are planning to use.

If the resistor was a regular metal oxide - then I agree it will sound considerably worse but if you invest in a nice sounding resistor like Caddock TF020 or Vishay bulk foil then I think you will have no problems at all - this was my experience.

I did not try tantalum yet but I guess they sound good to and perhaps the slightly cheaper caddock MK132 would be a slightly cheaper alternative - but I did not try that yet either.

mike

 

[TJF]

...and for my set up I had no problem.

Shure. It works also this way...

 

[jkeny]

I wonder how these Takman resistors would fare? Reasonable prices, I think, compared to some of the more exotic boutique parts http://www.diyhifisupply.com/?q=catalog/61/japanultrapremiumtakmanresistors
or here http://www.partsconnexion.com/t/catalog/resistors.html

 

[Tolu]

I have built the MKII version like George described. It has around 7k input impedance! It works great!

Then I did some research! After reading the data sheets and measuring charts of Silonex I thought that series R and shunt LDR is superior because of the signficant lower distortion.

That was theory, listenning is practice. Like always, don't trust paper! Don't judge anything before you did some serious listening!!!

I've built a DAC-controlled shunt section with LDRs and for series R I used some Vishay Dale CNY 55 10k Ohm resistors! Sound was a little bit flat and lifeless. So I used the theoretical worse LDR series/shunt configuration with a constant input impedance of 10k Ohm between -15 and -40 dB and around 50k down to -70dB.
The result was amazing.

The next step was source switching. I have read a lot about perfect signal switching and relays. I bought a handful of Meder reed relais and used a 1C version for signal switching. It was a loss in the audible signal that was remarkable. Then I used another LDR for signal/ source switching. What a difference. Perhaps it is subjective but with 3 LDRs in the signal path the sound is perfect. And I am very critical with this and imaging that I have golden ears (or at least silver)!

The big advantage of the RC-DAC-version is the easy-to-adjust impedance behavior of the whole setup.

 

[Bigun]

IF it turns out that series/shunt is better because of the quality of the resistor in the series position why not the following:

I build a variable shunt LDR volume control as I mentioned already. Instead of fixed resitors for the series resistance I use another LDR but have them (6 channels) at fixed LED brightness so I am using them as fixed resistors. I don't worry about matching them since I have trim pots on the shunt LDRs anyhow.

Now I have no loss of sound quality associated with series resistors that are not LDRs and yet don't have to match them. Of course I don't have a constant input impedance but I'm not too worries about that.

Better still, it means that I can use the LDRs that didn't yield good matching for the series resistors instead of letting them go to waste.

Better better still, the series/shunt arrangement will mean some automatic temperature compensation.

I don't want the complexity of a DAC - the application doesn't deserve it.

 

[ondesx]

I did an experiment suggested by someone through the thread. Since it's very easy to fix a resistance value for a given LDR, I used series LDRs with different fixed values through obviously different currents/voltages for each. The shunts remain "voltage controlled" by the 5 V/100 k pot. I tried 2K, 5K, 7K, 10 K. Of course, the usable range clearly changed in each case... I tried to compare more or less matched output power. My audio source is 100 ohms impedance with more than 4 V output signals and my amplified monitors are 10 K input impedance and have a sensitivity of 100 dB SPL for - 6 dB.

Believe it or not, the result remain better with "free" series LDRs ! I can't explain what is going on, but I noticed a degraded sound for "fixed" LDRs value !...

Any explanation welcome !...

Cheers,